103445 HEALTH EQUITY AND FINANCIAL PROTECTION REPORT TIMOR LESTE HEALTH EQUITY AND FINANCIAL PROTECTION REPORT TIMOR - LESTE ABOUT THE HEALTH EQUITY EALTH EQUITY AND FINANCIAL AND FINANCIAL PROTECTION REPORTS PROT ABOUT THE HEALTH EQUITY AND F The Health Equity and Financial al Protection Protection reports are reports short volumes are short country-specific that provide a picture country-spec on About The Health Equity and Financial Protection reports are in of equitythe the Health and Equity health financial and Financial protection in the Protection sectors health sectorsReports of low-of and middle-income low- and countries. Topics covered middle of equity include: inequalities utcomes, health and in health outcomes, financial behavior protection health behavior and health care utilization; and in health the health benefit incidence care sectors analysis; uti The Healthinclude: ogressivity financial Equity and Financial protection; andof inequalities Protection the reports are short progressivity health of health in country-specific care carehealthvolumes that financing. outcomes, provide financing. Data area picture drawn equity and of from thehealth financial behavio protection Data Demographic and are in Surveys,the financial health Health Surveys,sectors of low- Multiple and protection; middle-income World Health Surveys, Multiple countries. Topics Indicator and covered the include: inequalities Indicator Cluster Surveys,progressivity Cluster in health outcomes, Living Standards and of health Measurement care health behavior Surveys, and L utilization; benefit incidence health care Health Surveys, analysis; financial protection; and World the progressivity of health care financing. Data are drawn from the holdSurveys,surveys, as well as other household and surveys, and use a Health use a common common set of Surveys, health indicators Multiplefor setall countries of inIndicato the healt Demographic and Health Surveys,as Surveys, Health Surveys, World well as Multiple other Indicator Cluster Surveys, Living Standards household and Measurement Surveys, surveys, and as use a c series. using ted well All as other analyses household are conducted the surveys, health and use a using common setthe health of health modules modules indicators for allof the ADePT countries software. in the series. of Also the All analyses areavailable conductedADePT are Health using the Equity datasheets series. and Financial All Protection thatanalyses datasheetssummarize are conducted using the health mo health modules of the ADePT software. Also available arethat summarize Health key measures Equity and Financial key Protectionof equity datasheets and financial measures that summarize protection. key measures of of Equity equity and financial protection. and Financial Protection datasheets that summar The most recent Health Equity versions of the and Health Equity and Financial Protection reports Financial and datasheets can be downloaded Protection reports The yandhealth. The most recent most versions recent of the Health at www.worldbank.org/povertyandhealth. versions Equity and Financial of Protection reports the can and datasheets be downloadedEquity Health at and Financ www.worldbank.org/povertyandhealth . at www.worldbank.org/povertyandhealth. Full citation: World Bank. 2014. Health Equity and Financial Protection Report – Timor-Leste. Washington, D.C.: World Bank. Photo credit: Photo Alex Baluyut credit: Alex Baluyut ACKNOWLEDGEMENTS Health Equity and Financial Protection Report This report was produced and written by a task team consisting The authors would also like to thank Hans Beck (Senior of Caroline Ly (Consultant), Yi-Kyoung Lee (Senior Health Economist) and Martin Cumpa Castro (Consultant) for Specialist) and Xiaohui Hou (Senior Economist) under the task contributions and input on the data analysis, Sarah Harrison team leaders Yi-Kyoung Lee and Xiaohui Hou. The authors (Consultant) for contributions on the comparative analysis as thank Leander Buisman (Consultant), Caryn Bredenkamp well as Emiliana Gunawan (Program Assistant) and Daniela (Senior Economist) and Lourenco Camnahas (Consultant) Hoshino (Program Assistant) for administrative support, for their additional analysis, technical advice and/or written Devon Rohr (Consultant) for graphic design and Melody contributions. Toomas Palu (Sector Manager, Health, Nutrition Molinoff (Consultant) for proofreading. and Population of the East Asia and Pacific Region) provided The team would like to thank the Ministry of Health in Timor- technical comments and overall supervision on this report. The Leste for their support and the development partners for their team would also like to thank Franz Drees-Gross (Country feedbacks on the earlier drafts. Director for Timor-Leste, Papua New Guinea & Pacific Islands, East Asia and Pacific Region) and Luis Constantino (Country Financial support for this work was received from Rapid Social Manager for Timor-Leste) for their guidance and support. Response Multi-Donor Trust Fund. The report benefited from comments from Caryn Bredenkamp and Ajay Tandon (Senior Economist) on the earlier draft and peer reviews from Owen Smith (Senior Economist) and Tania Dmytraczenko (Senior Economist) and Augustine Asante (Research Fellow, Health Economics & Financing, UNSW). Timor-Leste 1 LIST OF ABBREVIATIONS AND ACRONYMS ARI Acute respiratory infection BIA Benefit-incidence analysis CHC Community health centers CI Concentration index CPI Consumer price index DHS Demographic and Health Survey EAP East Asia and Pacific GDP Gross domestic product GHE Government health expenditures GNI Gross national income HIES Household Income and Expenditures Survey HNGV Hospital Nacional Guido Valadares MCH Maternal and child health MICS Multiple Indicator Cluster Survey MoH Ministry of Health NHA National health accounts OOPs Out-of-Pocket Spending PPP Purchasing power parity SAMES Serviço Autónomo de Medicamentos e Equipamentos de Saúde SISCa Sistema Integradu Saude Communitaria THE Total health expenditure TLSLS Timor Leste Survey of Living Standard WDI World Development Indicators WHO World Health Organization 2 TABLE OF CONTENTS Executive Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1. Timor-Leste’s health system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1. Equity and financial protection as policy goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.2. Health financing system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.3. Health care delivery system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2. Inequalities in health. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1. Data availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2. Inequalities in health. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3. Inequalities in health care utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.1. Data availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.2. Inequalities in health care utilization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 4. Benefit incidence in government spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.1. Data availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.2. Inequalities in benefit incidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5. Financial protection in health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 5.1. Data availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 5.2. Catastrophic out-of-pocket payments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 5.3. Impoverishing out-of-pocket payments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 6. Progressivity of health financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 6.1. Data availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 6.2. Progressivity of health care financing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 7. Comparison with other countries on key equity measures. . . . . . . . . . . . . . . . . . . . . . . . . . 21 8. Summary and policy implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 8.1. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 8.2. Policy Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 9. References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 10. Annexes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 10.1. Measurement of indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 10.2. Methodological notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Timor-Leste 3 Health Equity and Financial Protection Report 4 EXECUTIVE SUMMARY This report analyses equity and financial protection in the health sector of Timor-Leste. In particular, it examines inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis; financial protection; and the progressivity of health care financing. Data are drawn from the 2009/2010 Demographic and Health Survey, the 2001-2002 and 2007-2008 Living Standards and Measurement Surveys as well as 2011-2012 Household Income Expenditure Survey, and the Ministry of Finance. All analyses are conducted using original data and performed using the health modules of the ADePT software. Is ill health more concentrated poorest two quintiles is only 5 percent, while their share of among the poor? consumption is 22 percent. Yes, with exceptions. In general, ill health is concentrated What are the policy implications? among the poor in Timor-Leste. This includes some selected 1. Continue improving the availability and quality of indicators of child health such as infant mortality, under-5 services at the frontline. The poor use community health mortality, stunting, and underweight. But wealthier households centers and mobile clinics more while the wealthy tend are more likely to report child illnesses such as diarrhea and to seek care from hospitals. The availability and quality fever, potentially because of their ability to more easily of services offered at the community health centers, recognize these symptoms. With respect to available measures health posts, and SISCA sites needs to be strengthened so of adult health, the better-off are more likely to experience patients, particularly the poor, can access quality primary sexually transmitted diseases among men, and obesity among care. Recently, a large number of Cuban trained Timorese non-pregnant women. Some risky health behaviors such as doctors have returned and present a unique opportunity smoking are more prevalent among the poor. to significantly strengthen the primary health care system, Do the poor use health services less especially in rural areas. This increase in the number of qualified health workers should be accompanied by than the rich? better management of human resources and the creation Yes. Health care utilization in Timor-Leste is concentrated of conducive environments for health care workers. The among the better-off. Immunizations, treatment of acute instruments include, but are not limited to: incentivizing respiratory infections, and use of bed nets are higher among good performance to reduce inefficiency such as total children from wealthier households. Skilled antenatal care and absenteeism and reduced working hours; providing basic delivery, contraceptive prevalence and use of bed nets are also infrastructure (e.g., water and electricity) and essential higher among women from wealthier households. commodities (e.g., medicines and medical supplies); and taking measures to improve quality (e.g., step-wise Is the distribution of government quality accreditation system). Engaging community and spending on health pro-rich or pro-poor? empowering citizens to effectively monitor service delivery will also be imperative to ensure effective utilization of the Pro-rich. Government spending on hospital care is found to services by citizens, particularly the poor. be pro-rich, regardless of the assumptions made in the benefit- incidence analysis. When it comes to lower level health care 2. Improve accessibility of secondary care to the poor services such as community health centers, health posts and when needed. The utilization of secondary care, which mobile clinics, government spending is pro-poor using two is largely publicly-subsidized, is pro-rich. In order for the of three methodological assumptions. Taken together, total poor to have better access to secondary care services, the up subsidies for health are found to be pro-rich. Generally, and down referral system between primary and secondary government expenditure on health favors the better-off. care levels needs to be strengthened. The major barriers for the poor to access secondary care might not be point What is the effect of out-of-pocket of service user fees, but rather physical access to district payments on household financial and national hospitals. The government needs to explore options which can improve the poor’s access to secondary well-being? care when clinically required by providing, for example, Minor. Out-of-pocket spending as a share of total health travel vouchers, or (partial) reimbursement of travel cost. expenditures is only about 4 percent in Timor-Leste. Only Anecdotal evidence has shown that oversea referrals are pro- 0.9 percent of households spend 10 percent of their household rich benefitting those who are economically and socially income or more on out-of-pocket health payments and 9.6 better off. Increased transparency on oversea referral percent of households spend 10 percent or more of nonfood programs is required to ensure the equitable allocation and consumption on out-of-pocket payments. Health spending has utilization of resources. a small effect on increasing the poverty rate. Less than 1 percent 3. Monitor the impact of tightening fiscal space on public of households fell under the $1.25/day poverty line because of health spending on poor. In the environment where fiscal health spending. This is similar to results from 2001/02 and space is tightening, the poor are usually more affected. The 2007/08. utilization of some essential health services that often incur Is health financing progressive out-of-pocket expenditure such as skilled delivery is much lower among the poor. This highlights the need for the or regressive? government to closely monitor the impact of the tightening Progressive. Taxes and out-of-pocket payments are also fiscal space on service utilization especially among the poor progressive relative to consumption. The wealthiest 20 percent as well as the catastrophic out-of-pocket expenditures. The cover 74 percent of total health care payments, compared to household surveys conducted by the MoH (e.g., DHS) and 39 percent of consumption. Health spending among the the MoF (e.g., TLSLS) can provide information to do so. Timor-Leste 5 1. TIMOR-LESTE’S HEALTH SYSTEM This section provides a brief overview of Timor-Leste’s health system, focusing on features that are likely to be especially salient for equity and financial protection. 1.1. Equity and financial protection as 1.2. Health financing system policy goals Health expenditure2 The Government of Timor-Leste is strongly committed to Timor-Leste spent US$ 82.1 (PPP) per capita on health or 5.1 improving financial protection in health and improving equity. percent of GDP in 20113. Government spending on health The following quote illustrates this commitment1: as a share of government expenditures has been declining in “The Mission of the Ministry of Health is to strive to recent years as a result of shifting government priorities. Health ensure the availability, accessibility and affordability spending as a share of total government spending declined of health services to all East Timorese people, to regulate from 12 percent in 2005 to an estimated 4 percent in 2012 the health sector and to promote community and , , . In 2011, this amounted to about US$ 58.1 (2011 PPP) 4 5 6 stakeholders participation...” or US$ 33.1(2011 current) per capita. External financing for the health sector, which may flow through government and Ministry of Health, Dili, Timor Leste private channels, increased from 38.2 percent of total health expenditures (THE) in 2007 to 50.8 percent in 20117. Table 1.1: Health expenditure data, 2011 Indicator Health expenditure as share of GDP 5.1% Government expenditure on health, per capita US$33.1 (Current) US$58.7 (PPP adjusted) Government expenditure on health as share of total health expenditures 71.5%7 Out-of-pocket (OOP) expenditure on health as share of total health expenditures 4.0% Source: WHO National Health Accounts database (2013) 1 Mission statement available at http://www.moh.gov.tl/?q=node/2, accessed 11/5/12 2 Data are from 2011 and available from the WHO Global Health Expenditure database, accessed 6/21/13 3 Ibid. 4 Timor-Leste Ministry of Finance Budget Portal, accessed 11/12/12. 5 Goods and Services as a share of total Goods and Services have dropped from about 13% to 4% between 2005 and 2012. 6 Without the human capital fund and infrastructure fund, the share of health is 6.4% and 6.2% in 2011 and 2012, respectively. 7 Note that external financing may be channelled through both the government and private sector. Thus, external financing and government financing may exceed 100%. Data are for 2011 and available from the WHO National Health Accounts database, accessed 6/21/13. 6 $!+  ##$#"#"275683 #!)#!%$-!"1"$!"$"  #   !- "# " ( #!), Health Equity and Financial  %!# " Protection $#Report ! <6,: !####"""&$!6,6,7566*!"!!=5,=!# !%!#!%$"!# #!$$, (:,5!#!%$" Centralization #'",  $ of and revenue-raising/sources ! "# funds !  #!$ !%$" " # # "#! %( Health%$! financing in #Timor-Leste #!$is ! highly centralized.  THE, $#$#" "$(,  '#! compared "$!" to 11.7 percent,  $ $# 7.0 percent and 0.1    percent =,!(*$(!%#!"!###"!%#$"$#--# Government spending accounted for 71.5 percent of total in PNG, Samoa and Tuvalu10. This includes OOPs on private spending as 2 "3 health (#" Figure shown in ! 1.1. &  In 2011, a large !" sector # #! health care. "#-" According to the 2007-2008 #  $#!" Timor-Leste >,  "  share or 80.8 percent of general government revenues came Survey of Living Standard (TLSLS), most of those who incurred !- "#$#!9,5!#*!#66,<!#*<,5!#5,6!# from the Petroleum Fund. Only 5.0 65 percent of revenues OOPs did so from visiting private sector. Most households  * came from taxes.  domestic Such$%$ ,  " heavy reliance $" " on petroleum (89 !%# percent) that "#! # sought health care !,  ! reported # # visiting a public 755<-755=!- "#$!%( %#!2  3*"##"&$!! " revenues makes the health sector vulnerable to petroleum price health care provider and only three percent of visits to the " ! fluctuations %"# and supply. !%# External "#!, sources make up  "# about a$"" half 2=> public !#3 provider ## incurred "$# a payment # to the ! provider 11 !!# . However, of THE %"#$#!!%!(#!!#%"#"##$!%!$!! 8 . Generally, publicly provided health care is free at households that visited private providers were more likely to the point of service and thus out-of-pocket (###!%!, payments (OOPs) pay OOPs for their services. More than half of patients who 66&%!*$""##%"#!%#!%!"&!!( are low in comparison to other East-Asia and the Pacific went to private providers made a payment to a provider12. # ( " ! #! "!%",  ! #   ##" & &# # !%# !%!"   countries9. OOPs in Timor-Leste account for 4.0 percent of (##!%!67,  Figure 1.1: Health care financing mix, 2007-2011         655? =5? ;5? $#-- # (#" #! !%#$!" 95? %!#$!" 75? 5? 755< 755= 755> 7565 7566  Source: WHO National Health Accounts database (2013)  ="#"##'#!!"$!"!#!$##$!%#!%$"$!", > $#-- #(#"!#((!#"#!$##!!#((#""$" "$!2 755;3%###+11&&&,&,#11#"1"##4 "4!%4,,   $"(#"!#!"!%"""!##",    ##$#""##"275683 66!# &"275653*#! $(!(#"#$#!&!!"$"!" #"!%"%!(&(!8!# !$#>;!# "#,!(##" "#$(""/(#"#%$"##$#!%!"*!"##! $#"(#"!!$!""##%!(##!"("#                              *0 67 !(##"&&##!%#!%!"!!##(#,"#$### ####"-2,,*3*###"$!%(.##$!##, 65 8 This indicates that external resources are channeled through both the public and private revenue sources. 9 Out-of-Pocket payments are typically direct spending after deducting third party payments such as insurance (WHO 2006) available at http://www.who. int/nha/methods/estimating_OOPs_ravi_final.pdf. The TLSMS includes payments for health care services and medicines as part of health spending. 10 WHO National Health Accounts estimates (2013) 11 According to Lewis (2010), the frequency of informal payments to public health care workers among users of health services vary widely from 3 percent in Peru to 96 percent in Pakistan. Informal payment in this study is defined as “payments to individual and institutional providers, in kind or in cash that are made outside official payment channels or are purchases meant to be covered by the health care system including value of medical supplies purchased by patients and drugs obtained from private pharmacies but intended to be part of government-financed health care services,” 12 Nearly half of patients who went to private providers reported to pay nothing. This might be due to the fact that patients paid in-kind (e.g., chicken), and that the survey didn’t capture that. Timor-Leste 7  Health Equity and Financial Protection Report !"&##!$#!"  Comparison ! with other &# countries #! $#!"  "# ") # "!  $  ! # %! # '#$!"!#%(!, "#*%!#'#$!""!$#" Compared with other countries in East Asia, the share of public financing for the overall health expenditure is relatively high in ! 71+7 government Timor-Leste: !# expenditure  !) "! as a #  share of THE -75+5 accounts for 71.7 !#.) percent in Timor,$# ! similar # to Thailand " (75.5 percent), but higher than Indonesia (34.1percent), Vietnam (40.4 percent) and Laos (49.3 percent). -34+1!#.)#-40+4!#. "-49+3!#.+  Figure 1.2: Government expenditure on health as a share of THE, 2011               80 75+5 71+5 70 60 49+3 50 45+7 40+4 -:. 40 34+1 33+3 30 22+4 20 13+0 10 0 Source: World Development Indicator (WDI, 2014) $!*!%# #!- )2014.  Government %!#spending for a large share accounts $#" " !of total health ! "! Timor (US$  ## 46) is only # slightly better '#$! than Myanmar $# (US$ !)  expenditure in Timor, but in absolute value, "$#%$)#$#" the amount is quite 23) and Laos (US$ 37) in parity purchasing power (PPP) term, $#"!##!$#!""#"+%!# small compared to other countries in East Asia. The government but still much lower than Thailand (US$ 202) which achieved "#!"$#""("#(!&"+ !#%!# needs to increase public health spending especially as the universal health coverage (Figure 1.3). #'#$!!- /46."("#(##!# (!- /23. "- / economy grows. Per capita government health expenditure in 37.  !#( $!" &! - . #!) $# "# $ &! #  - / 202. & %$%!"#%!-$!1+3.+ Figure 1.3: Government health expenditure per capita, 2011                0,, /02 +)%.,,1* /1, /,, .1, .,. .,, -1, 41 43 41 -,, 1- /3 02 1, ./ , Source: WDI (2014)   & ).,-0*  The public expenditure as a share of total government time public health spending accounted for 12percent of total  # expenditure is also low compared!# to other countries, only 2.9 government spending " in Timor. Yet, the argument to increase 11  % percent in Timor,$  .'4better slightly  % than Myanmar (1.3$ percent),   fiscal $ allocation on health is still*% )-'/  thin given the   projected limited )-0'1*'% far less than Thailand (14.5 percent). As noted earlier, this  government revenue (! growth in next few years, the competing figure might  "be 2'0 under-estimated   given the 2'. share of  health demand .,--  from .,-. other sectors " and the    need for   the health  sector    )2*'"!%$".,,1%"to was 6.4 percent and 6.2 percent in 2011 and 2012 without to show the effectiveness of resources spending in translation human capital funding and infrastructure funding (footnote 6). improved health outcomes.    this However,  is still  significantly lower than in 2005, -.   !   ' % at which             !    !! "#"$% 8       "  !        ! '              !    !! "#"$%       "  !   Health  Equity and   Financial Protection Report  ! '                       Figure 1.4: Public health expenditure as a share of government expenditure, 2011                            Source: WDI (2014)    ! ! !   ! !    !"!" #! +"  & ).,-0* /)3   "/)3,)  !   ! !    !!  %   !! ! Government spending on health!!$ ! ) !#!"!  11 !  %!"   0./. "!   !&   ! "! /4 !  0./0) Spending on the health sector mirrors national priorities on US$ 18 million in 2008 to US$ 14 million in 2012. There is   infrastructure 0.!0..6!13!0./0(! development (Figure 1.5A and Figure1.5B). "!"!  no clear explanation as to why goods and services have gone ! " Recent and forecast  health - 4  spending  0..6 !on is concentrated "!down. - /4 The  demand for  0./0) health " services !   is anticipated  to increase. !' !      #   expansion and rehabilitation of hospital capital works projects.      "! 37 !  0..6 -. ! 1. Therefore, goods and services, which account for a large part !0./0($!! Minor and major capital development accounted "!"! -/60..6! -/2 for 33 percent of the recurrent expenditures, should have increased. The expenditures in 2010%! of0./0)  but has recently decreased to about 16 !$&  observed # decrease #$)  raises concerns regarding the effectiveness of percent Salaries  in 2012.#  ! increased percent from 20!  !!  in 2008  at service delivery )  ' # of the frontline  services."! ' $ In 2012, transfers   to 35 percent in 2012; the absolute amount has more than accounted for 20 percent in the total budget, roughly US$ 10  !  ! "! %!" doubled from US$ 6 million in 2008 to about US$ 16 million ' " #  )   #      million per year. It raises serious equity concerns given a large in 2012. During this period!   the share of goods !# of time,  # and #& part of the ! ! ! transfers are for the  # oversea ) which treatment, 0./0'the ! has services  been decreasing   "! from0. ! about  ! 59 percent !! in 2008 to "!' "& better-off are most - /. likely  to  &)  !    benefit from. percent "!& 30 " in 2012; with the absolute#!!! amount dropping from  !# !!!'$! !!* !&!!) Figure 1.5A: Percent allocation of health budget items, 2008 - 2012             /..8 7.8 6.8 5.8 !#! 4.8 ! 3.8 2.8    1.8   #  0.8   /.8 .8 0..6 0..7 0./. 0.// 0./0   Timor-Leste 9 Health Equity and Financial Protection Report Figure 1.5B: Amount allocation of health budget items, 2008 – 2012 (US$, million)  #&       $""'$"#$!  50 45 40 35 "$" 30  " 25  ! ! 20 !  $! 15  ! 10 5 , 2008 2009 2010 2011 2012  Source: Ministry of Finance (2012) # * !" '.2012/   1.3.  Health care delivery system Provider organization Payment mechanisms and provider #%        The health sector is divided into four organizational levels: autonomy district health services, hospitals, Central services, and Health facilities are reliant on government financing allocated $  personalized (" services (i.e., (semi) autonomous agencies). by the central ministry level. Facilities have no significant  District health services are primarily responsible for delivering sources of internally generated funds (IGF). Only HNGV in  the " basic !"package. health service  ! $ " # They operate   (" 66 community $!* Dili charges  !" user for " fees" non-Timorese ! $!) or VIP !"!) patients and health " centers  ! $!)192 (CHCs),   posts, health mobile clinics !( ! $! as well .++) .!/ was able to #"#! !/+ keep about 70 percent of its IGF!" (MoH "2009). " The as the Sistema Integradu Saude Communitaria (SISCa). SISCa, ! $!   ' !! $ "!"! Ministry of Health (MoH) $+' makes personnel decisions, "66 limiting introduced in late 2007, is a community based program that the level of provider autonomy. #"'"" delivers !.!/)192"!"!)!!%!" preventative and curative health services, namely family            +   ) registration, mother and child care, nutrition monitoring " #  " 2007) !  #"' !    "" $ ! including  $""$ supplements, food # "$ curative care and " ! $!)Resource integrated ' ' availability !" ")and utilization "    control ""  )#" vector #!#"!)# and health promoting activities. The distribution "$ " "$" "  of" health posts around the country is designed so that every Taking into account the imminent inflow of new doctors, a ""$"!+!" #""!"! density of qualified #"#" '!!! health workers in Timor-Leste is getting citizen has a health facility within walking distance . Sparsely 13 "" $ ' "( !  " "' %" % close !" to the 13+   !' #"  !   benchmark recommended by WHO by the end of populated areas are served by mobile clinics and SISCas. ! $'!  !+!" The district health services work with the Central services ""! 2013 (i.e.,$!% 2.35/1000 %""" population) 15 . While ! about $! two-thirds of in"% implementing a wide range  of programs  !!#!#(") such as child the clinical health workers were   !+ nurses and midwives (n≅1500),  immunization, malaria and TB programs. the situation is changing very quickly with a major inflow of young medical graduates who have finished their training in !"!$  care through a" $" chain#$  !"!14"!" "! Hospitals deliver higher level health of Cuba and elsewhere (n≅790). The density of doctors per district "!". 14         five referral hospitals in the districts and one national hospital * /+!!%" (without incoming doctors) varies significantly from 0.04 to " " (Hospital " "Guido Nacional !" "   !" $HNGV) Valadares: " " in Dili. Cases !"! 0.35, while   that   of!!) " ! nurses (0.4-1.4) " and $ !! midwives  is (0.2-0.9) which cannot be treated at the district level are sent to the a little better. The addition of new health workers will not hospitals or in rare cases, transported overseas for treatment.  only increase the density of health professionals by 0.5/1,000 Central services provide centralized administrative support 13!"!!&## population, but +#"" it will also #" increase the )"  costs of service delivery. well as bulk purchasing as"!' services. Personalized services For example, it !"! !#")'"%40:50:"#"! is expected that drug prescription costs will include the Institute of Health Sciences that provides research increase by $2/person/year based on the prescription pattern of ""'%"# -!%+ !"!"%"!""% $" # !" " and in-service training, Serviço Autónomo de Medicamentos e skilled health workers. In addition, the MoH budgeted about !"+ Equipamentos de Saúde (SAMES; central medical store), and 14##) ) #) #!! #!! $12m to cover additional salaries and allowances, equipment, National Laboratory. housing, and training. 13 Walking distance is defined as a maximum of one hour. Due to the rough terrain in Timor, the bad conditions of many roads and the dispersed population, only between 40% and 50% of the population has a health facility within a one hour’s walk. The rest has to walk at least two or even three hours to reach a health post. 14 14 Baucau, Maliana, Suai, Oecussi and Maubesse 15 World Bank. Human Resources for Health in Timor-Leste: a Rapid Assessment, Nov 2012 10 2. INEQUALITIES IN HEALTH Most policymakers regard large inequalities in health outcomes between the poor and rich as undesirable. This section reports inequalities in child and adult health outcomes, as well as health behaviors. 2.1. Data availability A Demographic and Health Survey (DHS) was fielded in Timor- income measures, but one can construct an “asset index” using Leste from 2009 to 2010. The DHS has rich information for principal components analysis to rank households from poorest many health outcomes, particularly in relation to maternal to richest (see Filmer and Pritchett 2001). and child health (MCH). The DHS lacks consumption and 2.2. Inequalities in health The tables in this section show how health outcomes vary across Table 2.1 shows that infant and under-five mortality, stunting, asset (wealth) quintiles. The tables show the mean values of and underweight are worse among the poor. However, diarrhea, the indicator for each quintile, as well as for the sample as a acute respiratory infections (ARI) and fever are worse among whole. Also shown are the concentration indices (CI), which the better-off. The 2007/08 TLSLS confirms that diarrhea capture the direction and degree of inequality. A negative value and fever/malaria were self-reported at higher rates among indicates that the indicator takes higher values among the poor, the better-off. This may be because the better-off were better while a positive index indicates that the indicator takes higher able to identify the symptoms associated with these childhood values among the better-off. The larger the index in absolute diseases than the poor. Better educated households reported size, the more inequality there is. higher rates of diarrhea, fever and ARI than less educated households (data not shown). Table 2.1: Inequalities in child health, 2009/10 Child (<5 years) health Q1 Q2 Q3 Q4 Q5 Total CI Infant mortality rate (per 1,000 live births) 67.8 74.0 62.3 60.2 39.8 60.7 -0.094*** Under-five mortality rate (per 1,000 live births) 98.2 106.8 94.9 92.2 58.1 89.4 -0.092*** Stunting (%) 61.2 63.3 58.6 55.3 46.5 57.1 -0.054*** Underweight (%) 49.3 48.6 47.2 41.7 35.6 44.6 -0.065*** Diarrhea (%) 13.2 13.7 15.2 19.0 17.4 15.7 0.065*** Acute respiratory infection (%) 2.7 3.9 4.2 4.8 4.2 3.9 0.084*** Fever (%) 16.2 16.9 18.5 23.2 21.9 19.2 0.073*** Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%. Table 2.2 shows certain risk factors such as smoking are more concentrated among the poor for both men and women: Smoking rates among men are high at 70 percent, with higher prevalence of smoking concentrated among the poor. Smoking is not very prevalent among Timorese women, but higher among poor women. Less than one percent of non-pregnant women are obese. The prevalence of sexually transmitted diseases (STDs) among men is worse among the wealthy. Table 2.2: Inequalities in adult health and risk factors, 2009/10 Adult (15-49 years) health and risk factors Q1 Q2 Q3 Q4 Q5 Total CI Obesity among non-pregnant women (%) 0.5 0.6 0.6 1.0 1.6 0.9 0.269*** Anemia among women (%) 40.6 36.7 41.8 39.7 33.8 38.5 0.098*** Smoking among women (%) 6.7 5.1 4.5 3.7 3.5 4.6 -0.141*** Smoking among men (%) 73.3 73.9 70.4 68.5 61.3 69.5 -0.027*** STDs among men (%) 0.0 0.8 2.1 1.4 1.3 1.1 0.414*** Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest; Obesity: Body Mass Index >30 * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%. In sum, the tables in this section indicate that children from poor households in Timor-Leste are disproportionately affected by ill health. Adverse adult risk behaviors such as smoking are concentrated among poorer households. Timor-Leste 11 3. INEQUALITIES IN HEALTH CARE UTILIZATION In many countries, for a variety of possible reasons, the pattern of health care utilization tends to be distributed unequally across income groups, even after taking into account differences in medical needs. This section reports on inequalities in utilization of health care in Timor-Leste for different types of care, and for different types of health care providers. 3.1. Data availability As stated previously, a DHS was fielded in Timor-Leste from and income measures, but one can construct an “asset index” 2009 to 2010. The DHS has data on utilization of maternal using principal components analysis to rank households from and child health interventions. The DHS lacks consumption poorest to richest (see Filmer and Pritchett 2001). 3.2. Inequalities in health care utilization The tables in this section show how health utilization varies value indicates that the indicator takes higher values among the across consumption or asset quintiles. The tables show the poor, while a positive index indicates that the indicator takes mean values of the indicator for each quintile, as well as for the high values among the better-off. The larger the index in sample as a whole. Also shown are the concentration indices absolute size, the more inequality there is. which capture the direction and degree of inequality. A negative Table 3.1: Inequalities in maternal and child health interventions, 2009/10 Interventions Q1 Q2 Q3 Q4 Q5 Total CI Full immunization (%) 43.4 52.9 56.5 64.9 45.5 52.6 0.030** Treatment of diarrhea (%) 80.8 76.1 84.0 75.7 76.0 78.4 -0.011 Medical treatment of ARI (%) 53.2 66.8 75.6 73.2 73.6 69.6 0.054** Mosquito net use by children (%) 23.9 34.0 42.7 56.9 55.9 42.3 0.167*** Skilled antenatal care (4+ visits) (%) 41.1 44.9 57.3 63.6 68.8 55.1 0.112*** Skilled birth attendance (%) 10.6 14.2 21.2 38.9 69.5 30.3 0.392*** Contraceptive prevalence (%) 9.2 9.5 9.9 14.6 17.8 12.4 0.153*** Mosquito net use by pregnant women (%) 27.1 32.3 39.8 59.7 54.6 42.2 0.155*** Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%. Table 3.1 shows coverage of key MCH interventions, including care visits but only 30 percent of women deliver their baby the treatment of childhood illness is usually higher among the assisted by a skilled attendant. Women in the wealthiest quintile better-off. Around 53 percent of children are fully immunized, are 6.5 times more likely to deliver with a skilled attendant 70 percent of children with ARIs received medical treatment. than women in the poorest quintile. This is consistent with Use of mosquito nets by children (and pregnant women) is what the report will discuss later on utilization of secondary about 42 percent with children from the wealthiest quintile care by income quintiles. The poor usually do not seek care at twice more likely to use the nets than those from the poorest the secondary level. Another reason is that the fertility rate is quintile. Inequalities in medical treatment of diarrhea among quite high in Timor (5.7 on average according to DHS data). children under five are not statistically significant. About 55 Use of a modern method of contraception is more prevalent for percent of expectant women receive at least 4 skilled antenatal wealthy women. 12 4. BENEFIT INCIDENCE IN GOVERNMENT SPENDING Policymakers typically take the view that government health expenditure (GHE) ought not to disproportionately benefit the better-off, and if anything ought to favor the poor more than the better-off. Benefit-incidence analysis (BIA) shows whether and how far GHE disproportionately benefits the poor. This section reports BIA results for Timor-Leste, using three methods for allocating GHE to households, namely (i) the constant unit cost assumption, (ii) the constant unit subsidy assumption, and (iii) the proportional unit cost assumption. The first is arguably the least plausible of the three, since it implies that higher fees are not a reflection of more costly care (see Table 4.2). However, it does have the attraction of not needing to be modified if part of (general) GHE goes on demand-side subsidies through, for example, a conditional cash transfer for health. In Timor-Leste, recurrent spending on services is allocated to two major areas: district health services and hospitals. Generally, spending on hospitals is pro-rich whereas spending on district health services is pro-poor. Overall total government spending on health favors the rich. The results presented below were obtained using the constant-unit- subsidy and proportional-unit-cost assumptions; it is assumed implicitly that supply- and demand-side subsidies have the same distributional impact. 4.1. Data availability The two most recent Living Standards Measurement Surveys by provider category. However, the 2007/08 TLSLS provided (LSMS) for 2007/08 and 2011/12 (i.e., 2007/08 TLSLS and overlapping categories of public providers in their dataset: 2011/12 HIES) provide data on utilization of health services. doctor, nurse/paramedic, midwife in government health facility, Government expenditure data at the hospital and district mobile clinic, health posts, government community health levels from the MoH and Ministry of Finance (MoF) provide center, and government hospital. For the 2007/08 analysis, information on the amount of subsidies to public health care. providers were grouped into two categories: government hospitals and district visits. District-level visits are defined as all While the 2011/12 HIES provides more recent data, it has some other public visits to a government facility other than hospital limitations. The 2011/12 HIES has combined out-of-pocket (i.e. doctor, nurse/paramedic, midwife in government health spending data on public hospitals and clinics. The 2011/12 facility, mobile clinic, government sub-community health survey also does not collect data on the frequency of visits to center, and government community health center). health care providers nor does it separate out utilization data by whether or not a visit was made to a public or private facility. It Typically, a National Health Accounts (NHA) exercise provides instead asks whether or not an individual in a household visited detailed health financing data on various levels of care such as a provider in the last month. So there is potential bias in the sub-community health centers, health posts, outpatient hospital utilization data if an income group tended to visit health care care and inpatient hospital care. An NHA exercise has not been providers more frequently. Data on health care expenditures completed for Timor-Leste. Instead, the MoF’s 2010 Review would also not serve as an indicator of utilization since health of the Health Sector and the MoH’s 2009 update to the 2007 care is supposed to be free. Approximately three-fourths of all Medium Term Expenditure Framework (MTEF) provides households that reported a member sought health care also recurrent and capital expenditures for hospitals and districts. reported zero health care expenditure. For the purposes of this BIA, district recurrent expenditures are interpreted as total public outpatient subsidies and hospital On the other hand, the 2007/08 TLSLS captured frequency recurrent expenditures as total public hospital subsidies16. of data and separated out the category of health expenditures 4.2. Inequalities in benefit incidence The tables in this section show the distribution across with 0.333 visits per year per person on average. On the other consumption quintiles of utilization for government facilities, hand, hospitals, with an average of 0.236 visits per year per fees paid to these facilities, and estimated subsidies to the person, were less frequently used17. There is a striking contrast health sector. The latter depend on the assumptions made between the distribution of lower level health care services (sub- to allocate subsidies to households; results are presented for CHCs and mobile clinics) and use of general hospital services. three sets of assumptions. The tables show the shares of fees Utilization of the former steadily decreases with income. The or shares of subsidies that go to each quintile. Also shown are decrease is most noticeable for sub-CHCs, which drop from the concentration indices which capture the direction and an average number of 0.445 visits for the poorest quintile to degree of inequality. A negative value indicates that the variable 0.219 for the highest. In contrast, hospital visits increase with in question is higher among the poor, while a positive index levels of wealth. These patterns are reflected in their respective indicates higher values among the better off. The larger the concentration indices. The concentration index of sub-CHCs, index in absolute size, there is more inequality in the indicator. mobile clinics and CHCs are noticeably negative, indicating that use of this type of service is higher among the poor. By Table 4.1 shows the utilization of four types of public facilities contrast, the concentration index of hospital visits is positive, separated in 2007/08. According to the 2007/08 TLSLS, indicating that richer individuals use this service more than the the most frequently used type of service was the health post poor. The concentration index for CHCs was not statistically 16 Outpatient and inpatient subsidies are not distinguished in hospital subsidies. As a result, it is possible that we are underestimating outpatient subsidies because we do not account for the outpatient services delivered in the hospital setting. 17 Note: According to the MoF review of the health sector, district visits are 1.7 per capita and hospital visits are 0.2 per capita. Timor-Leste 13 Health Equity and Financial Protection Report significant. Data on utilization in 2011/12 is incomplete but In 2007/08, fees or OOPs paid to see public providers averaged analysis (not shown) reveals that the concentration index for all about US$ 0.4. This low mean is driven by a high number of public health facility visits was slightly negative, indicating that patients who paid nothing at the point of service. Among those the poor utilized public facilities more often than the wealthy. who did see public providers, 97 percent paid nothing. But among the patients who went to a public provider and reported Table 4.2 shows the inequalities in visits to individual health paying for services, they paid on average about US$ 11.7. care providers. The wealthy are more likely to see midwives as demonstrated by the positive concentration index. The concentration indices of visits to nurses/paramedics and doctors across indicators are not statistically significant. Table 4.1: Inequalities in use of public facilities (visits per capita per year), 2007/08 Quintile Sub-CHC Mobile Clinic CHC Hospital Poorest 0.445 0.134 0.174 0.179 2nd poorest 0.366 0.119 0.208 0.200 Median/middle 0.351 0.119 0.183 0.234 2nd wealthiest 0.288 0.102 0.172 0.255 Wealthiest 0.219 0.096 0.186 0.313 Total 0.333 0.114 0.185 0.236 Concentration Index -0.139*** -0.074** -0.003 0.118*** Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%. Table 4.2: Inequalities in use of public providers (visits per capita per year), 2007/08 Quintile Midwives Nurses/Paramedic Doctors Poorest 0.036 0.148 0.063 2nd poorest 0.039 0.168 0.060 Median/middle 0.054 0.131 0.065 2nd wealthiest 0.055 0.140 0.051 Wealthiest 0.068 0.125 0.063 Total 0.051 0.142 0.060 Concentration Index 0.137*** -0.040 -0.007 Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%. Table 4.3 shows the share of fees paid to district and hospital facilities but data on expenditures was disaggregated between facilities in 2007/08 and all public facilities in 2011/12. In public and private health facilities. Visits that were categorized 2007/08, visits to health care providers in the TLSLS were as public hospital and clinic facility visits include all reported categorized such that all visits to hospitals were categorized as visits minus visits associated with positive payments to a private public inpatient visits but could potentially include outpatient facility and visits to a traditional medical practitioner. It is visits. All other visits including visits to doctors, midwives possible that this utilization number still includes visits to a and nurses were categorized as outpatient visits but could private provider in which no payments were reported. Based potentially include visits to hospitals. In 2011/12, data on on this categorization, there appears to be a positive relationship utilization was not clearly separated between public and private between wealth and fees paid for public services. Table 4.3: Distribution of fees paid to public facility (%), 2007/08 & 2011/12 2007/08 2011/12 Quintile District level Hospital Hospital & Clinic Poorest 2.2 1.4 0.1 2nd poorest 11.2 0.8 3.3 Median/middle 22.3 14.8 6.3 2nd wealthiest 22.8 8.1 7.9 Wealthiest 41.6 74.9 82.5 Concentration Index 0.4079*** 0.6954*** 0.6989† Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS and 2011/12 HIES; District level facilities includes CHCs, sub-CHCs and mobile clinics Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%,† Standard error not computed; Data in 2007/08 reflect results after removing a household outlier. 14 Health Equity and Financial Protection Report Table 4.4 shows the incidence of government health spending. is slightly pro-rich, the overall pro-rich result is mainly due to The first two lines of the table show how aggregate government the effects of public spending in hospitals which are pro-rich. spending on health varies across hospital and district level When unit subsidies are assumed to be constant (the second services18. The table contains three sets of estimates of set of results), the subsidies become slightly more pro-rich with subsidy distribution across consumption quintiles. The first a concentration index of 0.1233. Finally, when unit costs are assumption is the constant unit cost assumption in which each assumed to be proportional to the amount spent OOP, the visit at a specific level of care cost the same amount irrespective subsidies for all types of services become considerably pro-rich. of the different services provided (e.g. the hospital inpatient The resulting picture is a strongly pro-rich incidence of public costs for an uncomplicated child delivery would be treated the spending that is driven by considerably pro-rich subsidies to same as the costs for a caesarean). This would be calculated by hospital services. taking the total costs of service (i.e. subsidies plus user fees) For 2011/12 data on total public health spending, we see that divided by the number of units of utilization. This method concentration indices for both the constant unit cost and unit however can result in negative imputed subsidies since the user subsidy assumptions are slightly negative but insignificant, fees paid for services may exceed the unit cost of service. To making payments somewhat proportional across quintiles. address this, any negative imputed subsidies are set to zero. However, the proportional cost assumption appears to be The second type of assumption is that the unit subsidy is pro-rich. The top two quintiles capture about 90 percent of constant, irrespective of the user fees. It is just calculated as government subsidies to the health sector. the total subsidy divided by the units of utilization. The third assumption is the proportional cost assumption. It assumes Taken together, these benefit incidence analyses find no that the size of subsidies is proportional to the size of user fees. absolute evidence that government spending on health favors This implies that the size of user fees reflects the case complexity the poor. On the contrary, if one assumes that higher fees also and therefore cost of services provided. reflect higher subsidies, one would conclude that government spending is, in fact, very much pro-rich and worsening when The first two lines indicate that 43 percent of recurrent compared to 2007/08. When using the other assumptions, one government subsidies are spent on public hospitals and 57 also finds that total subsidies favor the better-off in 2007/08 percent is spent at the district level. The first set of results and are proportional in 2011/12, an improvement from (based on the constant unit-cost assumption) for 2007/08 2007/08. Therefore, the assumptions are critical to interpreting shows that the poorest quintile receives on average 15.3 percent whether or not the 2011/12 data reflect a worsening in the of government health spending while the richest quintile benefit incidence. receives 26.1 percent. The corresponding concentration index (0.1111) is slightly positive. Although district-level utilization Table 4.4: Inequalities in the incidence of government health spending (%), 2007/08 & 2011/12 2007/08 2011/12 Districts Hospitals Total Total Total Subsidies (US$, million) 7.8 5.9 13.7 19.5 Share of Total Subsidy 56.9 43.1 100.0 100.0 Constant Unit Cost Assumption Poorest 18.5 12.9 15.3 18.4 2nd poorest 20.7 13.9 16.9 22.7 Median/middle 20.5 21.1 20.9 21.6 2nd wealthiest 20.5 21.1 20.8 14.7 Wealthiest 19.8 30.9 26.1 22.6 Total 100.0 100.0 100.0 100.0 Concentration Index 0.0328*** 0.2119*** 0.1111*** -0.0249 Constant Unit Subsidy Assumption Poorest 18.0 12.4 14.8 18.4 2nd poorest 20.5 13.5 16.5 22.6 Median/middle 20.4 21.1 20.8 21.6 2nd wealthiest 20.7 21.4 21.1 14.7 Wealthiest 20.4 31.6 26.8 22.7 Total 100.0 100.0 100.0 100.0 Concentration Index 0.0441*** 0.2248*** 0.1233*** -0.0237 Proportional Cost Assumption Poorest 2.2 1.4 1.7 0.1 2nd poorest 11.2 0.8 5.3 3.3 Median/middle 22.3 14.8 18.1 6.3 2nd wealthiest 22.8 8.1 14.4 7.9 Wealthiest 41.6 74.9 60.5 82.5 Total 100.0 100.0 100.0 100.0 Concentration Index 0.4079*** 0.6954*** 0.5504*** 0.7007*** Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS and 2011/12 HIES. The MoF’s 2010 Review of the Health Sector and the MoH’s 2009 update to the 2007 Medium Term Expenditure Framework (MTEF) provides recurrent and capital expenditures for hospitals and districts. Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%, † Standard error not computed; Data in 2007/08 reflect results after removing a household outlier. 18 It would be preferable to categorize the level of services as inpatient and outpatient services. However, the financial data from the MoH is reported as hospital and district services. Districts are expected to be responsible for primary care services while hospitals are responsible for inpatient services. However, it is possible that, without an effective gatekeeper system, care provided at hospitals may include outpatient services. Or, because of limitations in transport, districts may be providing inpatient services. Timor-Leste 15 5. FINANCIAL PROTECTION IN HEALTH Countries finance their health care through a mix of OOPs, private and social insurance, general revenues, and international development assistance. Therefore, the goal of health systems is not just to improve health but also to ensure that people are protected from the financial consequences of illness and death, or at least from the financial consequences of having to obtain medical care. This section presents data on two alternative measures of financial protection, one that considers whether OOPs is ‘catastrophic’, and the other that considers if it is ‘impoverishing’. Neither captures the income losses associated with illness, and both therefore underestimate the full financial impact of ill health on households. The section also explains the institutional arrangements used in Timor-Leste to provide financial protection in the health sector, and presents data on levels of inequalities in coverage. 5.1. Data availability Two LSMS were fielded in Timor-Leste from 2001 to 2002, provide data on health care spending as well as household total from 2007 to 2008, and in HIES from 2011 to 2012. They and non-food consumption. 5.2. Catastrophic out-of-pocket payments This subsection provides information on ‘catastrophic’ health The information in Table 5.1 on catastrophic payments is for payments. Catastrophic payments are defined as health care the 2001/02 TLSLS, 2007/08 TLSLS, and 2011/12 HIES. payments in excess of a predetermined percentage of the total According to the 2011/12 HIES data, when the threshold is household or nonfood spending. raised from 5 to 40 percent of total household expenditure, the estimate of the incidence of catastrophic payments falls from The columns of Table 5.1 give different thresholds above which 2.7 to 0 percent. However, using non-food expenditure, the health payment “budget shares” might be deemed catastrophic. estimate of the incidence of catastrophic payments falls from The first line of the table displays the catastrophic payment 13.4 to 2.7 percent. These results are similar to the 2007/08 “headcount”, i.e. the proportion of households with a health TLSLS and 2001/02 TLSLS, based on total consumption. payment budget share greater than the given threshold. The Based on non-food consumption the level of catastrophic second line relates the catastrophic payment headcount to payments in 2011/12 has improved. When the threshold is the household consumption distribution, and shows the raised from 5 to 40 percent of total household expenditure, the concentration index of the incidence of catastrophic payments. estimate of the incidence of catastrophic payments falls from a A positive value of the concentration index indicates a greater level of 3.7 percent to about 0 percent in the 2001/02 TLSLS; tendency for the better-off to have OOPs in excess of the it falls from 1.9 percent to about 0 percent in the 2007/08 payment threshold, whereas a negative value indicates that the TLSLS. Catastrophic payments are concentrated among the poor are more likely to have OOPs exceeding the threshold. rich for all thresholds based on total consumption in 2011. Table 5.1: Incidence of catastrophic out-of-pocket spending, 2001/02, 2007/08 & 2011/12 Threshold share of total household consumption 5% 10% 15% 25% 40% 2001/02 TLSLS Headcount 3.7 1.3 0.7 0.3 0.0 Concentration Index 0.043 0.061 -0.035 0.697*** 0.000† 2007/08 TLSLS Headcount 1.9 0.4 0.2 0.0 0.0 Concentration Index 0.244*** 0.155 0.212 0.965*** 0.000† 2011/12 HIES Headcount 2.7 0.9 0.4 0.2 0.0 Concentration Index 0.572*** 0.663*** 0.631*** 0.677*** 0.000† Threshold share of non-food consumption 5% 10% 15% 25% 40% 2001/02 TLSLS Headcount 25.6 22.7 20.7 17.1 14.2 Concentration Index 0.048 0.010 -0.036 -0.079** -0.121*** 2007/08 TLSLS Headcount 21.9 15.2 10.6 6.7 3.9 Concentration Index -0.024 -0.016 -0.055 -0.079* -0.022 2011/12 HIES Headcount 13.4 9.6 7.4 4.2 2.7 Concentration Index 0.401*** 0.365*** 0.376*** 0.382*** 0.400*** Source: Authors’ estimates using ADePT and data from the 2001-2002 and 2007-08 TLSLS; 2011/2012 HIES Notes: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%, † Standard error not computed. 16 Health Equity and Financial Protection Report 5.3. Impoverishing out-of-pocket payments This subsection presents poverty measures corresponding to poverty gap is obtained by simply dividing the poverty gap household consumption gross and net of OOPs. A comparison by the poverty line; this is useful when making comparisons of the two shows the scale of impoverishment due to health across countries with different poverty lines and currency units. payments. The idea is that a health problem necessitating Finally, the normalized mean positive poverty gap is a measure OOPs may be serious enough to push a household from being of the intensity of poverty, calculated by dividing the average above the poverty line before the health problem to being below poverty gap of the poor by the poverty line. the poverty line after the health problem. Adding OOPs to the Table 5.2 reports results for the 2001/02 TLSLS, the 2007/08 household’s nonmedical consumption (consumption including TLSLS and the 2011/12 HIES. Health expenditure has been – or gross of – health payments) gives us a sense of what the an insignificant driver of poverty: 22.5 percent, 35.7 percent, standard of living would have been without the health problem. and 44.1 percent of the population was poor using the $1.25/ The nonmedical spending (consumption excluding health Day line in 2001/02, 2007/08, and 2011/12, respectively. If we payments) gives us a sense of what the standard of living looks take OOPs out from the household’s consumption, recognizing like with the health problem. The assumption here is that OOPs that this expenditure is involuntary and simply enables a are involuntary and caused by health “shocks”; health spending household to cope with a health problem, the poverty rate goes is assumed to be financed by reducing current consumption. up to 22.9 percent, 36.1 percent, and 44.3 percent during the The first line of Table 5.2 shows the poverty “headcount” same period. Thus 0.4 percent of the population or less would which represents the proportion of population living below not have been poor if the resources they were forced to devote the poverty line at $1.25/day (PPP). The poverty gap is the to health care had been available to spend on other things. aggregate of all shortfalls from the poverty line. The normalized Table 5.2: Impoverishment through out-of-pocket health spending, 2001/02, 2007/08, & 2011/12 Consumption Consumption Absolute Percentage including excluding change change health payments health payments 2001/02 TLSLS Poverty Headcount 22.5% 22.9% 0.34pp 1.50% Poverty Gap $0.7 $0.7 $0.01 1.28% Normalized Poverty Gap $5.8 $5.9 $0.07 1.28% Normalized Mean Positive Poverty Gap $25.7 $25.9 $0.23 0.88% 2007/08 TLSLS Poverty Headcount 35.7% 36.1% 0.33pp 1.93% Poverty Gap $1.7 $1.7 $0.03 1.88% Normalized Poverty Gap $7.8 $7.9 $0.15 1.88% Normalized Mean Positive Poverty Gap $21.8 $22.0 $0.21 0.96% 2011/12 HIES Poverty Headcount 44.1% 44.3% 0.13pp 0.00% Poverty Gap $3.7 $3.7 $0.02 0.01% Normalized Poverty Gap $11.1 $11.2 $0.06 0.01% Normalized Mean Positive Poverty Gap $25.3 $25.3 $0.06 0.00% Source: Authors’ estimates using ADePT and data from the 2001/02, and 2007/08 TLSLSs and 2011/12 HIES. Note: Poverty line based on the $1.25/day line; pp: percentage point. Timor-Leste 17 Health Equity and Financial Protection Report Figure 5.1 effect  shows the  on poverty of OOPs ' via a “Pen’s In sum, the analyses '## in this section do not find high levels of !* parade”. Households are lined up in ascending order of their catastrophic health expenditure. To the extent that they are " ) consumption including OOPs. The vertical “paint drips” show present, catastrophic payments are found to be concentrated  to which OOPs divert a household’s spending away the extent among the wealthy19. With regards to impoverishment, the !' %   from items such as food, education, clothing, etc. The length data indicate that health  "  spending $ hardly increases !)  the absolute drip, of the paint $  %     shows therefore, how  '  far health   spending % number of   the !   impoverished.  Indeed, the    increase in the  compromises # %05)  living a household’s standards.  " '  poverty rate due to health  spending   is around 0.5 percent when %  !  !   ")  ' $1.25/day using the line.   " %   !    !/)2 #! .0)12+%)  Figure 5.1: The impoverishing effect of out-of-pocket spending               !(! *  ! 1/00+01  Source: Author’s ( " estimates using ADePT % and 2011/12 HIES .0)12+%)  Note: Poverty line based on the $1.25/day line.                !   %      !           !*  %  %'    " ,   !  !        !    -)  " " %      %   " %!'   !!)%   $!"% % "'   !*! !*!)  05 !!% '!  $   " # #%  % !    # %)$'  % "   ' ! &    % %  #' # %!$"") 1//3+/4! &  #    # %!,  %$"- "! %) 13 19 This could be a function of many possible factors, but the existing data do not provide concrete answers on why catastrophic payments are found to be concentrated among the wealthy. For example, a health financing system provides good financial protection for the poor, the utilization rates of the health system by the poor are low, or the wealthy are using more expensive services. The 2007/08 utilization data show that the wealthy do use (potentially more expensive) hospital services more frequently. 18 6. PROGRESSIVITY OF HEALTH FINANCING Health Equity and Financial Protection Report There is a general consensus that payments for health care ought to be at least proportional to households’ ability to pay, if not progressive (meaning a poor household contributes a smaller share of its resources than a rich one). The overall progressivity of a health financing system depends on the progressivity of each source of finance, and the share of health spending financed through each source. A system that relies exclusively on OOPs is likely to be regressive, since OOPs often absorbs a larger share of a poor household’s resources than of a rich household’s resources. This is not always the case, as it is likely that the poor are under-using health care, something that can be assessed by the distribution of health utilization. 6.1. Data availability The 2001/02 TLSLS, 2007/08 TLSLS and 2011/12 HIES to finance the health sector. It is assumed that individual’s tax provide data on household consumption, OOPs and tax contributions at the national level are proportional to the taxes payments by tax type. The MoF provides data on revenues for the health sector. for general government financing. There is no dedicated tax 6.2. Progressivity of health care financing The first five rows of Table 6.1 show each quintile’s average high degree of inequality with a Gini coefficient of 0.292 in consumption and financing share with households ranked 2011/12, which is comparable to 0.297 in 2007/08, and an in ascending order of gross consumption (i.e. consumption improvement from 0.403 in 2001/02. This indicates that the including health care payments). Health care payments are wealthy consume a large share of all the goods and services. For considered progressive if the poorest quintile’s share in total example, the wealthiest quintile consumes 38.7 percent of all household consumption exceeds its share in total payments, goods and services, almost as much as the poorest 60 percent while the opposite is true of the richest quintile. Payments are who consume 39.1 percent of all goods and services in 2011/12. regressive if the poorest quintile’s share in total consumption is For the health system to be progressive, the wealthy should be less than its share in total payments. Again, the opposite is true paying more than their share of total consumption in taxes and of the richest quintile. This exercise can be done for total health OOPs. In 2011, the Kakwani index was 0.385 indicating a care payments, as well as for each source separately. progressive health system. The wealthy paid 73.6 percent of the Table 6.1 shows the Gini coefficient, which measures the total taxes and OOPs received by the health system, compared degree of inequality in gross consumption—the higher the to 38.7 percent of consumption, making the health system number, the more unequal the distribution of consumption. progressive. This is an increase in progressivity from 2007/08 The line below that shows the concentration index, a measure which had an overall positive but not statistically significant of how unequally distributed health care payments are across Kakwani index of 0.045. In 2007/08, the wealthiest quintile consumption quintiles; a positive value indicates that payments consumed 39.1 percent of all goods and services, higher than are concentrated among the better off quintiles, while a negative the poorest 60 percent who consumed 38.8 percent of all goods index would indicate a concentration of payments among and services. In 2007/08, the health system was neutral to the poorer quintiles. The next line shows the Kakwani index, progressive, as the wealthy paid proportionate to their share defined as the concentration index less the Gini coefficient. A through both taxes and OOPs. positive value indicates that payments are more concentrated In 2011, each component of the public’s financing for the health among the better-off than consumption is and is a sign that system, taxes and OOPs were progressive too. The wealthy pay payments are progressive. A negative Kakwani index indicates more in taxes and OOPs and the distribution of the burden that payments are regressive. Finally, the table indicates the of taxes and OOPs is progressive relative to the distribution of size of the “redistributive effect” associated with health care consumption. The Kakwani index provides an indicator of the payments. This is the change in consumption inequality depth of progressivity. OOPs have a similar Kakwani index to brought about by health care payments. A positive number taxes which suggests that OOPs and taxes are both progressive. indicates that there was less inequality in consumption after payments than before, which is the case if payments are While taxes and OOPs constitute a small fraction of the total progressive. The more progressive they are, and the larger revenues to the health sector, it is still important to note the fraction of (gross) consumption accounted for health care that taxes and OOPs in Timor-Leste are progressive relative payments, the larger the amount of the “redistributive effect”. to consumption. Compared to previous years, the data in 2011/12 HIES show that the tax system and OOPs have Table 6.1 shows that health care financing in Timor-Leste has become significantly more progressive. Overall, payments to been progressive (i.e. health payments are progressive relative to the health system have become more progressive. consumption) in 2001/02, 2007/08 and 2011/12, as indicated by the positive Kakwani index results. Timor-Leste has a Timor-Leste 19 Health Equity and Financial Protection Report Table 6.1: Progressivity of health finance Consumption Taxes Out-of-Pocket Total Payments (from household) health Spending 2001/02 TLSLS Poorest 6.6 0.3 7.5 3.8 2nd poorest 10.6 0.3 4.2 2.2 Median/middle 14.5 0.3 15.4 7.7 2nd wealthiest 20.8 2.8 22.3 12.4 Wealthiest 47.6 96.3 50.5 73.8 Gini Coefficient 0.403*** Concentration Index 0.878*** 0.438*** 0.662*** Kakwani Index 0.475*** 0.035 0.259*** Redistributive Effect 0.4031 0.0012 0.0001 0.0013 2007/08 TLSLS Poorest 9.1 15.7 9.0 10.3 2nd poorest 13.0 1.3 10.5 8.7 Median/middle 16.7 6.2 19.0 16.4 2nd wealthiest 22.1 18.3 23.6 22.5 Wealthiest 39.1 58.4 38.0 42.1 Gini Coefficient 0.297*** Concentration Index 0.457*** 0.312*** 0.342*** Kakwani Index 0.160 0.016 0.045 Redistributive Effect 0.297 0.0003 0.0000 0.0003 2011/12 HIES Poorest 9.2 1.4 0.8 1.1 2nd poorest 13.2 4.2 3.6 3.9 Median/middle 16.8 4.5 6.2 5.4 2nd wealthiest 22.2 12.6 19.2 16.1 Wealthiest 38.7 77.4 70.2 73.6 Gini Coefficient 0.292*** Concentration Index 0.672*** 0.681*** 0.677*** Kakwani Index 0.380*** 0.389*** 0.385*** Redistributive Effect 0.2917 0.0032 0.0074 0.0105 Source: Distribution of consumption, taxes and OOPs estimated by authors using ADePT and data from the 2001/02 TLSLS, 2007/08 TLSLS and 2011/12 HIES. Notes: * Significant at 10%, **Significant at 5%, ***Significant at 1%. NHA weights for 2001/02 TLSLS applied 2.6% for OOPs and 2.7% for taxes. NHA weights for 2007/08 TSLS applied 2.6% for OOPs and 1.7% for taxes. NHA weights for 2011/12 HIES applied 4% for OOPs and 3.6% for taxes. Official Gini coefficient is 0.3193 for 2007/08. 20          7.  COMPARISON WITH   #  OTHER  ! COUNTRIES % ! ON   ! % KEY  EQUITY  MEASURES #      ! (     %  ' !*3   %'      -24 " .'     )  %   !" - . ! %" the This section will compare  equity measures of the four ) with key indicators    other comparator ! countries:  infant mortality   rate,  " under-5 & ) ! skilled mortality,  antenatal care! and (4+ visits), #  skilled  birth   attendance. ! These indicators  are closely  related to Millennium Development Goals (MDGs) on reducing child mortality '!"%%' and improving maternal health. It also reflects  ") both health outcomes and the health service utilization. Three countries were selected for comparison purposes based on  geographical closeness, similar survey years, and per capita GDP level.  "    !             %  !*3   %  '        ! )  "    $ -.  The !" variations   across the quintiles are smaller in Timor for both %!*3 both indicators as compared % countries. to other )  However, this infant mortality  "% !and under-5 mortality # rates, as compared   #to  “equitable” more  outcome ! is mainly  driven by   the / fact that    even the other three countries. Negative Concentration Index (CI) the better-off quintiles suffer        ! )  #"'   +! , !   % from higher infant mortality and indicates poor households have higher infant mortality and under-5 mortality rates. In general, this relates to the low living "%    "  *! ! under-5 mortality rates. Timor seems relatively equitable when   %!*3 standard and weakness in the overall health delivery system   % considering two   ) these  ' indicators because the     CI is closer  to 0 " in #in Timor.   #   "  "%%  )  Figure 7.1. Comparison with other countries on infant mortality rate (per 1000 live births)             *! ! )!        (! !!&   '! " !" &! # %! $ !"& $! % #! & !# "!    ! !#&             #!"! #!!( #!!) #!"! Source: Health Equity and Financial Protection Datasheets  !( ! %         !       Figure     7.2. Comparison with other countries on under-5 mortality rate (per 1,000 live births)       "#! ! "!! !!&       )! "       !" # '! $ !"& 10 %! % !# & #!    ! !#&             #!"! #!!( #!!) #!"! Source: Health Equity and Financial Protection Datasheets  & #        )                * 0'.+'  Timor-Leste 21  !    $  % %!!      #            '    !    !!%! #!!   & "!! !!#! Health Equity and Financial Protection Report   !! !!!!     #!!(   #!"!   #!"!   #!!) Timor’s antenatal care also seems more equitable than the other However, huge inequality exists in Timor in terms of skilled  countries (Figure 7.3). The positive CIs show that the antenatal  !          birth attendance (Figure 7.4). Compared with other countries, utilization  is in favor of the rich, that is, women who are better Timor-Leste has the highest CI at the value of 0.4, showing off are more likely to receive antenatal care in all four countries. the skilled birth attendance is strongly in favor of the rich. The                                  Compared with the three other countries, the distribution of    results confirm the earlier findings          that the bottom quintiles this indicator seems slightly more equitable in Timor-Leste, were much less likely to use secondary care.              because   *!! *!!*!! the absolute*!! CI is closer to zero. !"'! !"'! !"'!!"'!     !     !     !     ! )!! )!! )!!)!! with other countries Figure 7.3. Comparison  on skilled antenatal care (4+ visits) !"%! !"%! !"%! !"%! "#! (!! (!! (!! (!! !%& !"#! !"#! !"#! !"#!         '!! '!! '!! '!! !% " " " " "!! !"!! !"!! !"!! !"!! &!! &!! &!! &!! !$& # # # # !!)! !!)! !!)! !!)! )! %!!%!! %!! %!! !$ " $ $ $ $ !#& !!'! !!'! !!'! !!'! # $!! $!! $!!$!!        '! % % % % !# !!%! !!%! $ !!%! !!%! #!! #!! #!!#!! % & & & & %! !!#! !"& !!#! !!#! !!#! "!! "!! "!!"!! &         !" !! #! !!!! !! !!!! !!!! !!!! !!!!             #!"!   #!!(   #!"!   #!"!   #!"!   #!!(   #!!(   #!!(   #!!)   #!!)   #!!)   #!!) !!&   #!"!   #!"!   #!"!   #!"!   !  !           !           !           !          !                 #!"! #!!( #!!) #!"! Source: Health Equity and Financial Protection Datasheets                                                                !                  Figure    7.4. Comparison with other countries on percent of birth attended by skilled health workers    *!!  !"'! "#! "#! "#! "#!               !%&  !%& !%& !%&        #%     ! )!! !"%!     "       !% !% !% !% "!! "!! (!! "!! "!! !$& !"#! !$& !$& !$&   '!! )! )! )! )! !$ !$ !$ !$ " "" " " !"!! $$ &!! !#& !#& !#&!#& # # # # #                             '! '! '! '! !!)! %!! !# !#!# !# $ $ $ $ $ !!'! $!! %! %! %!%! !"& !"& !"&!"& % % % % % !!%! & & & & #!! !" !" !" !" #! #! #! #! & !!#!             "!! !!& !!& !!&!!&   ! !!! !! ! !!! !!!!                 #!"!               #!!(                  #!!)           #!"! #!"! #!"! #!"! #!!( #!"! #!!( #!!( #!!) #!!( #!!) #!"! #!!) #!!) #!"! #!"! #!"!   !                 Source: Health Equity and Financial Protection Datasheets   !           !           !           !          In summary,     compared with other countries, the distribution equitably distributed across the quintiles in Timor than other              of infant mortality and under-5 mortality rates seems more countries, skilled birth attendance is significantly pro-rich.       equitable in Timor. However, this is largely driven by the fact However, even with the much-improved access of skilled birth  that     even        the     better-off      households             suffer   from a     high    relatively            attendance        by the richest        quintile,       health   the   #%  #%     outcomes  #%  #% (infant      "     "     "     " infant and under-5 mortality rate. This relates to the low living                         and under-5 mortality) of those are still quite low compared in general and weakness in the overall health delivery standard "#! with the other three countries, implying quality of the service !%& system in Timor. Though the antenatal visits seem more remains as an issue. !% "!! $$ $$ $$ $$ 22 !$& )! !$ " 8. SUMMARY AND POLICY IMPLICATIONS 8.1. Summary Timor-Leste has achieved relatively high levels of financial The poor tend to utilize lower level health care services such as protection for its citizens, but there remain some concerns. community health centers and mobile clinics more frequently While the government finances a large share of total health than the wealthy. The wealthy use hospitals more frequently expenditure (72%) compared to other East Asian countries, which are more costly than lower level health care services. If Timor-Leste still relies heavily on external financing, an one assumes that higher fees also reflect higher subsidies, one unstable source of financing that flows through public and would conclude that government spending is, in fact, very private channels. It accounts for almost half of total health much pro‐rich. expenditure. Publicly provided health care is free at the point Health care spending is a small share of household expenditures of service, making out-of-pocket spending a small share of total and has little effect on driving households into poverty. With health expenditure. out-of-pocket spending accounting for a small share of total Government spending accounts for a large share (72%) of the health expenditures, household health spending is not a major total health expenditure, which is relatively high compared driver of poverty. In 2011, less than ten percent of households with other East Asia countries at similar or higher GDP levels. spent ten percent of non-food consumption on health, However, in terms of absolute value of government spending compared to 23 percent in 2001/02. Also, less than one percent on health or public health expenditure as a share of government of households face the additional risk of impoverishment due to expenditure, Timor is among the lowest of the comparator health expenses throughout the same period. countries. Public spending has been steadily decreasing from Timor-Leste has a high degree of inequality, but health care 12 percent of the GDP in 2005, down to an estimated four financing in Timor-Leste is progressive. The Gini coefficient percent in 2012. Also, the share of goods and services has been was 0.292 in 2011/12, comparable to 0.297 in 2007/08, and decreasing in recent years from about 59 percent of the health an improvement from 0.403 in 2001/02. This indicates that budget in 2008 to 30 percent in 2012. This could potentially the wealthy consume a large share of all the goods and services exacerbate health outcomes for the poor and the country’s (39 percent). The wealthy, however, also cover 74 percent of ability to limit catastrophic health spending. total health care payment, and thus health care financing is The poor tend to have worse child health outcomes, compared progressive. While taxes and OOPs constitute a small fraction to wealthier households. Infants and children under five of the total revenues to the health sector, it is still important from the households in the poorest quintile are 1.4 times to note that taxes and OOPs in Timor-Leste are progressive more likely to be underweight and 1.7 times more likely to relative to consumption. die than those from households in the wealthiest quintile. Compared with Indonesia, Philippines, and Cambodia, Wealthier households are more likely to report child illnesses the distribution of infant mortality and under-5 mortality such as diarrhea and fever. This may be due to the ability, rate seems more equitable in Timor. However, this is largely among the wealthy, to more easily recognize the symptoms of driven by the fact that even the better-off households suffer diseases, even though they may actually have a lower incidence from a relatively high infant and under-5 mortality. Though of these illnesses. Coverage of key maternal and child health the antenatal visits seem more equitably distributed across the interventions including treatment of childhood illness is quintiles in Timor, the inequality in skilled birth attendance usually higher among the better-off. Children from the indicates this service is significantly pro-rich. However, even wealthiest households, for example, are twice more likely to use with the much improved access of skilled birth attendance bednets than those from the poorest households. Women in by the richest quintile, the health outcomes are still quite low the wealthiest households are 6.5 times more likely to deliver compared with the other three countries, implying quality of with a skilled attendant than women in the poorest households. the service remains as an issue. Timor-Leste 23 Health Equity and Financial Protection Report 8.2. Policy Implications 1. Continue improving the availability and quality of 2. Improve accessibility of secondary care to the poor services at the frontline. As established, the poor access when needed. The utilization of secondary care, which community health centers and mobile clinics more often, is largely publicly-subsidized, is pro-rich. In order for the while the wealthy tend to seek care from hospitals. The poor to have better access to secondary care services, the up availability and quality of services offered at the community and down referral system between primary and secondary health centers, health posts, and SISCA sites needs to be care levels needs to be strengthened. The major barriers strengthened so patients, particularly the poor, can access for the poor to access secondary care might not be point quality primary care. A large number of Cuban trained of service user fees, but rather physical access to district Timorese doctors have recently returned and present a and national hospitals. The government needs to explore unique opportunity to significantly strengthen the primary options which can improve the poor’s access to secondary health care system, especially in rural areas. The increase care when clinically required by providing, for example, in the number of qualified health workers should be travel vouchers, or (partial) reimbursement of travel costs. accompanied by better management of human resources Anecdotal evidence has shown that oversea referrals are pro- and creating conducive environments for health care rich benefitting those who are economically and socially workers. The instruments include but are not limited to: better off. Increased transparency on the oversea referral incentivizing good performance to reduce inefficiency program is required to ensure the equitable allocation and such as total absenteeism and reduced working hours; utilization of resources. improving procurement and timely distribution of essential 3. Monitor the impact of tightening fiscal space on public commodities (e.g., medicines and medical supplies); health spending on poor. In an environment where fiscal ensuring basic infrastructure (e.g., water and electricity); space is tightening, the poor are usually more affected. The and taking measures to improve quality of services utilization of some essential health services that often incur (e.g., step-wise quality accreditation system). Engaging out-of-pocket expenditure such as skilled delivery is much community and empowering citizens to effectively monitor lower among the poor. This highlights the need for the service delivery will also be imperative to ensure effective government to closely monitor the impact of the tightening utilization of the services by citizens, particularly the poor. fiscal space on service utilization especially among the poor, as well as the catastrophic out-of-pocket expenditures. The household surveys conducted by the MoH (e.g., DHS) and the MoF (e.g., TLSLS) provide information to do so. 24 9. REFERENCES Eden, HV. Bessette, F., Pedastsaar, E and Nayer, H. (2010). Public Financial Management Report. Washington, D.C.: International Monetary Fund. Filmer, D and L. Pritchett (2001). “Estimating wealth effects without expenditure data or tears: An application to educational enrollments in states of India.” Demography 38 (1): 115-132. Gwatkin, D., S. Rutstein, K. Johnson, E. Suliman, A. Wagstaff and A. Amouzou. (2007). Socioeconomic differences in Health, Nutrition and Population. Washington, DC: World Bank. Lewis M (2010). Informal Payments and the Financing of Health Care in Developing and Transition Countries. IPC Working Paper Series Number 98 O’Donnell, O., E. van Doorslaer, A. Wagstaff and M. Lindelow. (2008). Analyzing health equity using household survey data: a guide to techniques and their implementation. Washington, D.C: World Bank. Timor Leste Ministry of Finance. (2010). Expenditure Review of the Health Sector. Dili: Ministry of Finance. Timor Leste Ministry of Finance. (2012). Budget Transparency Portal. Dili: Ministry of Finance. Available at http://www.mof.gov.tl/?lang=en. Timor Leste Ministry of Health. (2009). Medium Term Expenditure Framework. Dili: Ministry of Health, Government of Timor Leste. Wagstaff, A. (2012). “Benefit-incidence analysis: are government health expenditures more pro-rich than we think?” Health Economics 21(4): 351-66. Wagstaff, A., M. Bilger, Z. Sajaia and M. Lokshin. (2011). Health equity and financial protection: streamlined analysis with ADePT software. Washington, D.C.: World Bank. World Health Organization. (2012). Global Health Observatory: National Health Accounts. Estimates for Timor-Leste. Geneva: WHO. World Bank. (2012). World Development Indicators. Washington, D.C.: World Bank. Timor-Leste 25 10. ANNEXES 10.1 Measurement of indicators Indicator Measurement Data Child Health Infant mortality rate Number of deaths among children under 12 months of age per 1,000 live births DHS Under-five mortality rate Number of deaths among children under 5 years of age per 1,000 live births DHS Stunting % of children with a height-for-age z-score <-2 standard deviations from the reference DHS median (Note: z-score calculated using WHO 2006 Child Growth Standards) Underweight % of children with a weight-for-age z-score <-2 standard deviations from the reference DHS median (Note: z-score calculated using WHO 2006 Child Growth Standards) Diarrhea % of children with diarrhea (past two weeks) DHS Acute respiratory infection % of children with an episode of coughing and rapid breathing (past two weeks) DHS Fever % of children with fever (past two weeks) DHS Adult Health Obesity among non- % of women aged 15 to 49 with a BMI above 30 DHS pregnant women Anemia % of women aged 15 to 49 anemic based on hemoglobin testing DHS Risk Factors Smoking (all) % of adults who smoke any tobacco products such as cigarettes, cigars or pipes DHS Smoking (women) % of women aged 15 to 49 who smoke cigarettes, pipe or other tobacco DHS Smoking (men) % of women aged 18 to 49 who smoke cigarettes, pipe or other tobacco DHS Mosquito net use % of children who slept under an (ever) insecticide treated bed net (ITN) (past night) DHS by children Mosquito net use by % of pregnant women aged 15 to 49 who slept under an (ever) insecticide treated DHS pregnant women bed net (ITN) (past night) Maternal and Child Health Interventions Full immunization % of children aged 12-23 months who received BCG, measles, and three doses DHS, MICS of polio and DPT, either verified by card or by recall of respondent Treatment of diarrhea % of children with diarrhea given oral rehydration salts (ORS) or home-made solution DHS, MICS Medical treatment of ARI % of children with a cough and rapid breathing who sought medical treatment for DHS, MICS acute respiratory infection (past 2 weeks) Skilled antenatal care % of mothers aged 15 to 49 who received at least 4 antenatal care visits from any DHS (4+ visits) skilled personnel (doctor, nurse/midwife, auxiliary midwife, family nurse, trained birth attendant) Skilled birth attendance % of mothers aged 15 to 49 that were attended by any skilled personnel at child’s birth DHS Contraceptive prevalence % of women aged 15 to 49 who currently use a modern method of contraception DHS, MICS 26 Health Equity and Financial Protection Report 10.2. Methodological notes Sections 2 and 3: Inequalities in health and health Differences between the child mortality statistics in this report care utilization and in the DHS are due to the use of a “true cohort life table approach” which gives the true probabilities of deaths. This The selection and measurement of health outcome indicators means that we define infant mortality as the number of deaths used in Section 2 and 3 on inequalities in health and health among children under 12 months of age per 1,000 live births, care utilization was based on (i) a comparison of indicators used using a sample of children born between 10 and 1 years before in major health publications and databases, (ii) the advice of the survey, by dropping the children born in the year before World Bank Health Specialists on recommended monitoring the survey (i.e. children who have not had the chance to and measurement practice in their respective fields, and (iii) complete one year). Similarly, we define under-five mortality how measurable those indicators would be in the available as the number of deaths among children under 60 months data sources. The following major reports/databases were of age per 1,000 live births, using a sample of children born consulted as a guide to indicator measurement: World Bank between 10 and 5 years before the survey, by dropping the Development Indicators, the World Bank’s HNPStats database, children born in the 60 months preceding the survey from the WHO’s World Health Survey country reports, and the World sample (i.e. children who have not had the chance to complete Bank’s report series on “Socio-economic differences in health, five years). By contrast, DHS adopts a “synthetic cohort life nutrition and population (Gwatkin et al. 2007). table approach”. This approach involves creating a number of The data sources for this section include the Demographic and age segments and calculating the probability of dying for each Health Surveys (DHS) and multipurpose household surveys segment. The denominators are the number of children in the (such as the World Bank Living Standard and Measurement cohort who turned exactly 0, 1, 3, 6, 12, 24, 36, 48 months Surveys). Where the selected indicators are available in more in the period 0-10 years before the survey and have completed than one of these surveys, all measures are reported. the respective age segments in the period of 0-10 years before the survey. The numerator of each one is the number of deaths In all analyses of inequality in this section, i.e. quintile analysis for each group of children in the respective age range (e.g. and calculation of concentration indices, households are ranked 0-0.99 months, 1-2.99 months, 3-5.99 months, etc). Once the by an asset index computed using principal components analysis. probability of each has been calculated, they are combined to In order to avoid presenting estimates biased by insufficient calculate the IMR and U5MR using the lifetable function of power, indicators were removed from the tables if the sample “q”(=probability of dying). Yet another alternative is a “vital size in any quintile was less than the following thresholds: 250 statistics approach” where the denominator is births 0-10 years per quintile for infant and child mortality estimates and 25 per before the survey and the numerator is all the under- 5 deaths quintile for all other indicators. This follows the practice of in the period 0-10 years before the survey. This means not Gwatkin et al. (2007). In addition, the statistical significance of only the deaths of the children who were born in 0-10 years all concentration indices is reported. before the survey, but also the children who were born in 10- 15 years before the survey (but died under the age of 5 in the period of 0-10 years before the survey). The DHS approach is probably technically superior because the information for the denominator and numerator come from exactly the same children, and is not affected by a rapid change in the number of births (as the vital statistics approach is) nor does it require dropping information on children born in recent years (as the true cohort life table approach does). It is, however, more computationally-intensive and, as such, a less pragmatic one for our purposes. The approach used in the Socioeconomic Differences reports is also a “true cohort life table approach”, except that in the case of under-five mortality these reports also only drop children born in the last year (rather than the last five years) from the sample. Timor-Leste 27 Health Equity and Financial Protection Report Section 4: Benefit-incidence analysis The section on benefit incidence analysis uses three different methods for allocating government health expenditure to households, invoking three different assumptions that are described in detail in Wagstaff (2011). The first, the constant unit cost assumption, treats the sum of individual fees and government subsidies as constant, and thus any fees paid when using public services results in a reduction in the government subsidy received. The second, the constant unit subsidy assumption, allocates the same subsidy to each unit of service used, irrespective of the fees paid. Finally, the third, the proportional unit cost assumption, makes the cost of care proportional to the fees paid, which implies that the government subsidy received increases as the fees paid increases. In calculating the distribution of fees, service utilization and government subsidies, households are ranked by per capita consumption. The quintile data sources for this section include multipurpose household surveys that are used to obtain information on service utilization at difference levels of care and fees paid by patients. Data on government subsidies at each level of service are obtained from government reports on budgets and health expenditures. Section 5: Financial protection This section examines catastrophic health care payments and impoverishment due to OOPs. In this section, households are ranked by consumption. The analysis of catastrophic health care payments follows the popular approach elaborated upon O’Donnell et al. (2008) which defines health spending as “catastrophic” if it exceeds some fraction or threshold of total expenditure, or of total nonfood expenditure, in a given period. As O’Donnell et al. (2008) note, the threshold of 10% for total expenditure and 40% for nonfood expenditure are commonly used in the literature. In addition to measures of incidence, distribution-sensitive measures of catastrophic payments are calculated, specifically the concentration index, and statistical significance is reported. The analysis of impoverishing expenditure uses the national poverty lines. Section 6: Progressivity of health care finance This section examines the progressivity of different sources of healthcare financing/payments, including OOPs, health insurance contributions, direct taxation and indirect taxation. The Kakwani index, defined as the concentration index minus the Gini coefficient, indicates whether payments are more/less concentrated among the better off than consumption is and, thus, is a sign of whether payments are progressive/regressive. The main data source needed for the analysis of progressivity of health care financing is a multipurpose household survey, preferably with a very detailed consumption module. 28 TECTION REPORTS FINANCIAL PROTECTION REPORTS cific volumes that provide a picture e short country-specific e-income countries. volumes T that provide opics a picture covered of low- and middle-income countries. ilization; benefit incidence analysis; Topics covered or and health care utilization; e drawn from the Demographic and benefit incidence analysis; e financing. Living Data are drawn Standards andfrom the Demographic and Measurement or Cluster Surveys, Living th indicators for all countriesStandards and Measurement in the common set of health indicators software. Also available are for all countries Health in the odules equity of the andADePT Also available are Health software. protection. financial rize key measures of equity and financial protection. s and datasheets can be downloaded cial Protection reports and datasheets can be downloaded Photo credit: Alex Baluyut Photo credit: Alex Baluyut EALTH EQUITY AND ABOUT THE HEALTH EQUITY FINANCIAL AND FINANCIAL PROTECTION REPORTS PROT ABOUT THE HEALTH EQUITY AND F The Health Equity and Financial al Protection Protection reports are reports short volumes are short country-specific that provide a picture country-spec in on of equityThe the Health and financial protection health Equity in thesectors health sectors Financial and of low-ofand middle-income low-Protection and countries. reports Topics covered middle are of equity include: inequalities utcomes, health and financial in health outcomes,behavior protection health behavior and health care utilization; and in the health health care benefit incidence sectors analysis; uti ogressivity financialinclude: protection; and ofinequalities the health progressivity in of health health care care financing. outcomes, financing. Data are drawn from the health Data Demographic behavio and are financial Health Surveys, Surveys, Multiple protection; World Health Surveys, Multiple Indicator and the progressivity Cluster Indicator Cluster Surveys, Living Standards and of health Measurement care Surveys, L Surveys, hold Health surveys, Surveys, as well as other household and World surveys,use and use a common Health a common Surveys, Multiple set of health indicators for set all countries of Indicato the in healt ted series. using Surveys, All analyses the well as health are conducted using as other the health modules moduleshousehold of of the ADePT software. surveys, the Also availableand are Health ADePT use a c datasheets series. Allthat Equity and Financial Protection analyses summarize datasheets are conducted that summarize key measureskey using of equity and the measures health financial protection. mo of Equity and Financial Protection datasheets that summar The most recent Health Equityversions of theand Health Equity Financial Protection and Financial Protection reports reports and datasheets can be downloaded yandhealth.The most recent versions of the Health Equity and Financ at www.worldbank.org/povertyandhealth. at www.worldbank.org/povertyandhealth. Photo credit: Alex Baluyut